In this paper, we present MOSAIIC, an agent-based model to simulate the road traffic of a city in the context of a catastrophic event. Whether natural (cyclone, earthquake, flood) or human (industrial accident) in origin, catastrophic situations modify both infrastructures (buildings, road networks) and human behaviors, which can have a huge impact on human safety. Because the heterogeneities of human behaviors, of land-uses and of network topology have a great impact on the traffic flows, the agent-based modeling is particularly adapted to this subject. In this paper, we focus on the new traffic model itself: the way geographical data is used to build a network, the various behaviors of our agents, from the individual to the collective level. IntroductionNowadays, traffic simulations are often used by urban planners to make decisions concerning road infrastructures. Many models have been developed these last years. These models are grouped according to their levels of representation: macroscopic A modeling approach that is particularly well-fitted for micro-simulation is agent-based modeling. It allows to consider the heterogeneity of driver behaviors and to take into account the global impact of local processes.Such approach is increasingly used as many frameworks allowing urban planners to easily build their own scenarios (MATSIM [5], SUMO [6], AgentPolis [7]) are

State-of-the-art methods for handwriting text recognition are based on deep learning approaches and language modeling that require large data sets during training. In practice, there are some applications where the system processes mono-writer documents, and would thus benefit from being trained on examples from that writer. However, this is not common to have numerous examples coming from just one writer. In this paper, we propose an approach to adapt both the optical model and the language model to a particular writer, from a generic system trained on large data sets with a variety of examples. We show the benefits of the optical and language model writer adaptation. Our approach reaches competitive results on the READ 2018 data set, which is dedicated to model adaptation to particular writers.

No abstract